SSIINNGGLLEE NNUUCCLLEEOOTTIIDDEE PPOOLLYYMMOORRPPHHIISSMM AANNAALLYYSSIISS IINN

AAPPPPLLIICCAATTIIOONN TTOO FFIINNEE GGEENNEE MMAAPPPPIINNGG

by

Manish Sampat Pungliya

A Thesis submitted to the Faculty of the

WORCESTER POLYTECHNIC INSTITUTE

in partial fulfillment of the requirements for the

Degree of Master of Science

In

Biology

By

___________________

May 2001

APPROVED BY:

Dr. Julia Krushkal , Major Advisor Dr. Elizabeth Ryder, Advisor on Record

Dr. Carolina Ruiz, Thesis Committee Dr. David Adams, Thesis Committee

Dr. Ronald Cheetham, Head of the Department

ii

ABSTRACT

Single nucleotide polymorphisms (SNPs) are single base variations among

groups of individuals. In order to study their properties in fine gene mapping, I

considered their occurrence as transitions and transversions. The aim of the study was

to classify each polymorphism depending upon whether it was a transition or

transversion and to calculate the proportions of transitions and transversions in the

SNP data from the public databases. This ratio was found to be 2.35 for data from the

Whitehead Institute for Genome Research database, 2.003 from the Genome

Database, and 2.086 from the SNP Consortium database. These results indicate that

the ratio of the numbers of transitions to transversions was very different than the

expected ratio of 0.5. To study the effect of different transition to transversion ratios

in fine gene mapping, a simulation study was performed to generate nucleotide

sequence data. The study investigated the effect of different transition to transversion

ratios on linkage disequilibrium parameter (LD), which is frequently used in

association analysis to identify functional mutations. My results showed no

considerable effect of different transition to transversion ratios on LD. I also studied

the distribution of allele frequencies of biallelic SNPs from the Genome Database.

My results showed that the most common SNPs are normally distributed with mean

allele frequency of 0.7520 and standard deviation of 0.1272. These results can be

useful in future studies for simulating SNP behavior. I also studied the simulated data

provided by the Genetic Analysis Workshop 12 to identify functional SNPs in

candidate genes by using the genotype-specific linkage disequilibrium method.

iii

ACKNOWLEDGMENTS

I am very thankful to many people who were directly and indirectly involved

in my completion of my thesis. This is the only way to express my sincere gratitude

towards them. First of all I would like to thank my advisor Dr. Julia Krushkal for

giving me the unique opportunity of working in the exciting field of Bioinformatics.

Her continuous encouragement and support made this thesis as one of the most

important things to be cherished throughout my life.

I extend my special gratitude to Dr. Elizabeth Ryder, Dr. David Adams and

Dr. Carolina Ruiz for their sincere suggestions and helpful discussions throughout my

years of graduate study. Special thanks goes to Dr. Matthew Ward for providing me

with the program ‘Scansort’. I am also thankful to Christopher Shoemaker and

Michael Sao Pedro who did the data conversion and provided me with the formatted

data for Genetic Analysis Workshop 12 (GAW12). I would also like to make a

mention of my friend Raju Subramanian for guiding me in writing algorithms in C++.

Analysis of the GAW12 data presented in this thesis was funded by the grant

“Computational algorithms for analysis of genomic data” from the Research

Development Council of Worcester Polytechnic Institute. I am also thankful to the

National Institutes of Health grant GM31575 to GAW12.

Finally, I express my gratitude to my parents for their unconditional love and

moral support throughout my graduate study. And last but not the least I would like to

thank all my friends at WPI who have made my stay here a memorable thing in my

life.

iv

TABLE OF CONTENTS

Page no.

Abstract………………………………………………………………………………..ii

Acknowledgments..…………………………………………………………………..iii

List of figures and tables……………………………………………………………....v

Introduction………..……………………………………………………………….….1

Thesis objectives……………………………………………………………………..13

Part I: Investigation of the effect of transition to transversion ratio Methods………………………………………………………………………………14 Results………………………………………………………………………………..28

Discussion……………………………………………………………………………38 Part II: Application of SNPs in fine gene mapping Introduction……………….………………………………………………………….41 Methods………………………………………………………………………………45

Results………………………………………………………………………………..50

Discussion……………………………………………………………………………56

Bibliography………………………………………………………………………....60

Appendices …………………………………………………………………………..64

v

LIST OF FIGURES AND TABLES

Page no.

Figure 1: Linkage disequilibrium phenomena………………………………………………05

Figure 2: One-parameter model (Jukes and Cantor 1969)…………………………………..10

Figure 3: Two-parameter model (Kimura 1980)…………………………………………….12

Figure 4: Simulation of population history by TREEVOLVE………………………………19

Figure 5: Sequence simulation by TREEVOLVE after the population history has been simulated………………………………………………………...22

Figure 6: An example of input parameter file for TREEVOLVE……………………………22

Table 1: Input parameter values for TREEVOLVE used in this study………………………23

Figure 7: A sample output of TREEVOLVE for sample size of 7 with sequence length

of 43..………………………………………………………………………………..24

Table 2: The number of simulations performed……………………………………………...24

Figure 8: Sequence.arp (An input file for Arlequin for the calculation of LD)……………...26

Figure 9: Batch.arb…………………………………………………………………………...27

Table 3: Data summary from all the three databases………………………………………...28

Figure 10: SNP distribution in the chromosome 6 data from the Whitehead

Institute of Genome Research…...…………………………………………………..29

Figure 11: SNP distribution in the chromosome 6 data from the Genome Database……….30

Figure 12: SNP distribution in the chromosome 6 data from the SNP Consortium ………...31

Table 4: Summarized results from figures 10, 11, and 12……………………………………31

Figure 13: Allele frequency distribution for all common alleles of SNPs from human chromosome 6………...……………………………………………………..33

Table 5: Mean linkage disequilibrium values ± standard deviation for different

values of the transition to transversion ratios at different recombination rates……...34

Figure 14: Plot of Mean LD Vs. α/2β at recombination rate (r) equal to 0.0………………..36

Figure 15: Plot of Mean LD Vs. α/2β at recombination rate (r) equal to 10-8……………….36

Figure 16: Plot of Mean LD Vs. α/2β at recombination rate (r) equal to 3×10-5…………….37

Figure 17: The phenotypic model of the data simulated by GAW12………………………...44

Figure 18: Founders in a pedigree……………………………………………………………45

Table 6: Calculation of maximum difference by Scansort program…………………………48

Table 7: Chi-square analysis for an individual SNP position………………………………..49

Table 8: Scansort output for data set 1 (8250 pedigree founders) for top 20

SNPs sorted by the value of d…………...…………………………………………..50

vi

Table 9: The 20 most significant SNPs in gene 1. The data set analyzed was

the 8250 pedigree founders from all the 50 replicates………………………...…….51

Table 10: The 15 most significant SNPs in gene 2. The data set analyzed was

the 8250 pedigree founders from all the 50 replicates…………………...………….52

Table 11: The 12 most significant SNPs in gene 6. The data set analyzed

was the 8250 pedigree founders from all the 50 replicates…………………...……..52

Table 12. Number of significant sequence polymorphisms, the range of their

significance values, and the total number of polymorphisms, by gene, determined

from the 8250 pedigree founders from all the 50 replicates……………………...….53

Table 13: Number of significant SNPs in genes 1 and 2 analyzed separately for 8250

pedigree founders……………………………………………………………...…….54

Table 14: Number of significant SNPs in genes 1 and 2 analyzed separately for 165

pedigree founders in best replicate 42……………………………………...………..54

Table 15: Number of significant SNPs in genes 1 and 2 analyzed separately

for 1000 individuals in best replicate 42………………….…………………………55

1

INTRODUCTION

At the present time, we are at a stage where we can read nearly the entire

genetic code of the human genome, as recently a rough draft of the human genome

sequence has been determined (International Human Genome Sequencing

Consortium 2001; Venter et. al 2001). It represents the sequence of A, G, C, and T

letters that are symbols for nucleotides. The specific sequence of these nucleotides

constitutes all our genes that have specific characteristics and expression in the

human being. These genes are responsible for various physical, physiological and

pharmacological activities in the body. Since mutations in these genes are often the

cause of many heritable diseases, it is sometimes necessary to find specific genetic

mutations responsible for a particular disease. This is one of the current aims of the

Human Genome Project.

Until now, genome analysts have concentrated their efforts on finding the

similarities between different individuals, and they have found that 99.9 percent of

anyone’s genes perfectly match those of another individual (Brown 2000). But the

remaining 0.1 percent of the genes varies among individuals. It is these nucleotide

variations that are of interest to many researchers, as these variations might change

the properties of tha t particular gene. Even a simple single nucleotide polymorphism

(SNP) in a gene sequence can change the property of that gene and cause a disease, as

differences in DNA of that gene could change the phenotype. Below I describe the

properties of SNPs in more detail.

2

Single Nucleotide Polymorphism

SNPs are the most common single base variations in the human population. A

SNP is a variation where two alternative bases occur at appreciable frequency (the

frequency of each base is above 1% in a population) (Lander et. al 1998). Most of

these variants are neutral, but some are functional. One of the important goals of

genetic analysis is to identify those SNPs and SNP variants (alleles), which are

associated with a disease. For example, consider the following nucleotide sequence,

AATTTCCGG

AATTACCGG

The change from T to A is considered a SNP provided both alleles are present in

more than 1 percent in the population. SNPs are often binary, i.e. they most often

have only two alleles. They are less susceptible to mutations than microsatellite

repeat markers. A microsatellite is a short sequence of repeated nucleotides in a

genome e.g.: AATGAATGAATG-----, where AATG is repeated a variable number of

times. Due to their stability, SNPs are very useful for studying human evolutionary

history.

Importance of SNPs

The occurrence of SNPs is approximately one in every 1000-2000 base pairs, and

the total number of SNPs in the human genome estimated by November 2000 is

1,433,393 (The International SNP Map Working Group 2001) and 2,104,820 (Venter

3

et. al 2001). These polymorphisms are present in coding as well non-coding regions.

Less than 1% of all the SNPs are present in the protein-coding regions of the genome

(Venter et. al 2001). This suggests that a very small proportion of SNPs may be

responsible for phenotypic variation.

SNPs in association analysis

Association analysis and linkage analysis are methods for gene mapping of

complex human disorders. In the case of association analysis, one tries to find an

association between a marker locus (could be a SNP) and a disease locus (may be or

may not be a SNP) from the genetic data of the human population (Risch and

Merikangas 1996). Such analysis tests for association of loci at short distances apart

and attempts to identify individual mutations. SNPs are therefore very useful in

association analysis, as SNPs are so frequent and close to each other that the loci stay

associated even after much recombination. In contrast, in linkage analysis, one tests

for genetic linkage between a disease locus and a marker locus which is generally a

microsatellite marker (Kruglyak et al. 1996). The distances detected in linkage

analysis are generally much larger than those in association analysis. In linkage

studies, one needs to have a family with a certain proportion of individuals with the

disease. Linkage analysis takes into account all the available information from a

pedigree of a chromosomal region and a trait that cosegregate. Such sample data are

often difficult to obtain. This disadvantage is overcome by using association analysis,

as it does not require the pedigree information for a trait.

4

Association analysis does not involve the creation of a model based on the

pedigree information. In fact, it is based on the association that exists between two

phenotypes or loci when they occur together in a group of individuals more often than

expected by chance. Association may or may not be due to linkage, as in linkage one

tries to find an association between two loci in a pedigree.

Association analysis is a very useful tool in mapping genes for complex

phenotypes. An etiology of a complex phenotype can be associated with factors such

as marker locus, disease gene, other disease genes, environment, and cultural factors.

All of these factors interact with each other in a very complex way, resulting in the

expression of that phenotype. The main purpose of an association study to map

disease genes and find causative mutations is to minimize the effect of other genetic,

environmental, and cultural factors, and in turn increase the correlations of marker

and disease gene with the complex phenotype. By increasing the number of marker

loci, one can improve the efficacy of association studies in finding functional

mutations, as there is always an increased probability of finding at least one marker

locus associated with the disease gene.

The disadvantage of a linkage study compared to an association study lies in its

basic requirement of number of affected individuals necessary to detect the linkage to

a complex trait (Risch and Merikangas 1996). This number is very large when the

proportion of affected individuals in a population is small. In contrast, the

requirement of such number is vastly less in association methods as one tries to study

association between a single locus or multiple loci together with the disease locus in

affected/unaffected individuals. One would have better understanding of the

5

association of a disease locus with other loci that lie on the same chromosome or on

different chromosomes when genome-wide association tests are performed.

Linkage disequilibrium

To investigate the association between the two loci/multiple loci, one of the

ways of evaluating association is to use a parameter called linkage disequilibrium.

Linkage disequilibrium mapping is a frequently used analytical approach involving

SNPs. Linkage disequilibrium (LD) is defined as a nonrandom association between

SNPs in proximity to each other (Terwilliger and Weiss 1998). So if thousands of

SNPs are mapped over the entire genome, then through LD, associations could be

established between any of the susceptibility regions of the gene and a particular SNP

marker. Two sites are said to be in linkage disequilibrium if the presence of one

marker locus at one site enhances the predictability of another locus on the same

chromosome or different chromosome. This indicates that these two sites are

associated, and it helps in mapping the disease gene, as one of the loci could be the

disease locus, or might be associated with the disease locus so that it is transmitted in

a family.

(A,a) (B,b) locus 1 locus 2 AB Ab aB ab Figure 1: Linkage disequilibrium phenomena (see text for explanation)

6

Suppose there are two loci, 1 and 2, that are close to each other (Figure 1). Loci 1

and 2 may be on the same or on different chromosomes with two alleles each: A,a

and B,b respective ly. As a result, there are four haplotypes possible: AB, Ab, aB, and

ab. If allele A has a frequency of ρA in the population and allele B has a frequency of

ρB, then haplotype AB would have frequency ρAρB in the absence of association, as

they would occur independent of each other in the population. If alleles A and B are

associated, then the frequency of haplotype AB would be ρAρB + D, where D is the

measure of the strength of LD between the two loci 1 and 2.

D = ρAB - ρAρB

If allele B at locus 2 is a disease causing locus, then the frequency of allele A would

be much higher in affected individuals than in unaffected individuals because of its

association with B. This method of association analysis is used in case-control

studies. In case-control studies, there are two samples, one of cases (with disease) and

one of control (no disease) individuals. The two groups are further classified

according to one marker on the basis of its presence or absence. By performing chi-

square analysis, one can estimate the significance of the association of the marker

with the disease. This method can be extrapolated to markers having more than one

allele (reviewed by Elston 1998).

Thus if one has a large map of such marker loci over the entire genome, one can

test them to find the linkage disequilibrium with the disease locus of interest. The

markers are generally polymorphisms like microsatellites, or single nucleotide

polymorphisms within or outside the gene, but lying close to it so that they could be

associated with the disease locus. This is the basis for linkage disequilibrium mapping

7

of common disease genes. Linkage disequilibrium mapping is considered as the

indirect strategy (Kruglyak 1999) of association studies, as it involves testing for

associations between the disease locus and nearby polymorphism which is not

physically linked to the disease gene. It is done by using a dense map of polymorphic

markers across the genome which can be tested for association with the disease locus.

Biallelic single nucleotide polymorphisms are the markers of choice for this kind of

analysis because of their abundance in the genome and low mutation rates compared

to microsatellites. This is particularly important for whole-genome association studies

for identification of complex disease genes (Schafer and Hawkins, 1998). Lai et al

(1998) have proved the feasibility and importance of creating such SNP maps for

identification of genes of interest. The importance of creating a dense map of SNPs is

that it is useful in linkage disequilibrium mapping so that a susceptibility allele and a

marker lie within the range of linkage disequilibrium (McCarthy and Hilfiker 2000).

In the second half of this thesis, I have presented the application of association

analysis to a simulated data set.

Recent ly, it has been shown that the average extent of linkage disequilibrium in

the general human population is approximately 3kb; thus roughly 500,000 SNPs (as

markers) may be needed for systematic whole-genome linkage disequilibrium studies

(Kruglyak 1999) so that one would have good placing of SNP markers (in the range

of 3kb) over the entire genome. LD tends to decrease when the distance between the

markers is in the range of 10-100kb, as there is a high probability of recombination

and genetic drift with increased distance (McCarthy and Hilfiker 2000).

8

Application of SNPs in Pharmacogenomics

Linkage disequilibrium mapping has also been important in pharmacogenomic

studies (McCarthy and Hilfiker 2000). The use of LD mapping using SNPs has

gained a lot of importance, as it provides the necessary information about the drug

response in a genetically heterogeneous population used in clinical trials. It will help

to uncover the secret of why some drugs are effective in certain people and not in

others. These small variations may cause differences in drug response among patients

as they alter gene expression. Some drugs may have a beneficial effect on some

patients, but might prove harmful to others. So in the future a simple genetic test may

determine whether an individual can be treated effectively by a given drug. Such

application of genomics to pharmaceuticals is categorized as Pharmacogenomics,

which is the branch of genomics addressing molecular pharmacology and toxicology.

Hopefully this new branch will help reduce the cost of drug development (presently

it is in the range of $400-500 million), at the same time increasing the speed of the

development process (Rothberg, Ramesh and Burgess 2000). Pharmacogenomics

involves detecting and cataloguing SNPs in genes responsible for drug response. This

will uncover the variability of individual drug responses, and facilitate appropriate

patient selection during clinical trials and the aftermarket of a particular drug.

Other applications

SNPs are present in any part of the human genome. If they are present in the

coding or regulatory part of the gene, a SNP variant might alter gene function, and

9

thus may be related to the disease progression. However, if a SNP is not functional it

is still important in mapping studies, as it might be present very close to the disease

gene. This is an important basis of linkage disequilibrium analysis as one can study

the presence of that particular SNP in a population to find out its association with the

disease gene in the population.

Presence of SNPs

The presence of SNPs is related to nucleotide substitutions in DNA sequences.

To study the process of nucleotide substitutions, several mathematical models based

on the probability of nucleotide substitutions have been proposed in the literature

(reviewed by Li 1997). The two simplest and most frequently used models are

1) Jukes and Cantor’s (1969) one-parameter model and

2) Kimura’s (1980) two-parameter model

10

1) Jukes and Cantor’s one-parameter model

Purines A G

Pyrimidines C T

Fig.2: One -parameter model (Jukes and Cantor 1969)

This model assumes the substitution of nucleotides in DNA sequences is a

random process with all possible changes occurring with equal probabilities. For

example, nucleotide A can change to nucleotides T or C or G with equal probability.

Let α be the rate of substitution per time in each of the three possible directions of

nucleotide A. In this model, the rate of substitution for each nucleotide is 3α. Because

only one parameter (α) is involved in this model, it is called a one-parameter model.

The basic assumption of the one-parameter model (that all nucleotide

substitutions occur randomly) is unrealistic in most cases. For example, transitions

(changes between A and G, or between C and T) are generally more frequent than

transversions (changes between A and C or T, and between G and C or T, and vice

versa). To take this fact into account, a two-parameter model was proposed by

Kimura (1980).

αα

αα

αα

αα αα αα

11

2) Kimura’s (1980) two-parameter model

This model attempts to account for different frequencies of transitions and

transversions (Figure 3). In this model, there are two parameters involved. One is α,

which is the rate of transition (changes within purines or pyrimidines) and the other is

β , which is the rate of transversion (changes between purines and pyrimidines). If

transitions and transversions occur with equal rates (α = β), the expected ratio

between all possible transitions and all possible transversions from the two-parameter

model would be,

Ntransitions/Ntransversions = α/2β = 0.5

as there are four possibilities of transitions and eight possibilities of transversions.

Figure 3: Two-parameter model (Kimura 1980)

One- and two-parameter models of nucleotide substitutions may not apply to a

relatively short evolutionary time and also the possibility of occurring four types of

transitions (or eight types of transversions) with the same rate may not be true. There

Purines A G

Pyrimidines C T

αα

ββ

αα

ββ ββ ββ

12

are many other substitution models, some of which were described by Li (1997).

These models are more complicated mathematically and take into account complex

nucleotide substitution matrix, as some of these higher models are based on six-

parameters or nine-parameters substitution matrix (as these models consider different

rates of substitutions within transitions or transversions). Felsenstein’s (1981)

suggested the equal- input model, which is based on the equal rate of substitution of

one nucleotide with the other three nucleotides (similar to the one-parameter model).

Hasegawa et al. (1985) suggested a newer model (Model HKY85) with the addition

of additional substitution parameters and base frequencies. All these models are

mathematically complicated to compute. Felsenstein’s, Jukes and Cantor’s and

Kimura’s models are special cases of Hasegawa’s model. Felsenstein’s model takes

into account the ratio of transition to transversion equal to 1.0. Jukes and Cantor’s

model (one-parameter model) considers equal nucleotide frequencies and transition to

transversion ratio to 1.0, while Kimura’s model (two-parameter model) considers

only equal nucleotide frequencies.

13

THESIS OBJECTIVES

In the first part of my work, I investigated the exact ratio of transition to

transversion in SNPs from the publicly available databases and compared it with the

expected ratio of 0.5 (α = β). I also studied the distribution of the most common

SNPs from one of the data sets explained in the next section.

In previous studies, most association methods that used SNPs did not take into

account the possible differences between the rates of transitions and transversions.

These studies considered the transition and transversion occurring with equal rates as

in the one-parameter model. To investigate the validity of this approach, after finding

the actual proportion of transitions/transversions in the public databases, I studied the

effect of different transition to transversion ratios (including observed and expected)

on fine gene mapping using computer simulations.

In the second part of this thesis, I describe how association methods are

applied for fine gene mapping using simulated SNP data from Genetic Analysis

Workshop 12 (GAW 12).

Part I

Investigation of the effect of the transition to

transversion ratio

14

METHODS

DATA COLLECTION

To find out the ratio of transition to transversion in the SNP data, the SNPs

from human chromosome 6 were collected. I selected human chromosome 6

specifically because it has a number of genes associated with human diseases such as

breast cancer, hypertension (Krushkal et al. 1999), maple syrup urine disease

(Nobukuni et al. 1991), diabetes mellitus (Davies et al. 1994), psoriasis (Balendran et

al. 1999), and schizophrenia (Cao et al. 1997)1. I used three publicly available

databases to collect the data:

1) Whitehead Institute of Biomedical Research/MIT Center for Genome

Research2

2) The Genome Database 3

3) The SNP Consortium Ltd 4

1) Whitehead Institute of Biomedical Research/MIT Center for Genome

Research

This center started work on SNPs with the intention of developing a

dense map of SNPs (about 100,000 in number) (Collins, Guyer, and

Chakravarti 1997). This database has an anonymous list of SNPs from both

coding and noncoding regions of the genome. The SNPs are listed on a

1 http://www.ncbi.nlm.nih.gov:80/htbin-post/Omim/getmap?d2417 2 http://www-genome.wi.mit.edu 3 http://www.gdb.org 4 http://snp.cshl.org

15

genetic distance map (cM) with corresponding sequence tagged sites (STSs)

(Lander et al. 1998). The output is shown in Appendix 1.

2) The Genome Database (GDB)

This database is a result of an international collaboration in support of

the Human Genome Project. It has the list of genes for each chromosome. In

this database, the list of SNPs, labeled as point variations, is provided for

individual genes. Another advantage of the GDB is that it also provides the

allele frequencies for many polymorphisms, which is useful in finding the

allele frequency distribution for SNPs in human population. The data

collected from this database are in Appendix 2.

3) The SNP Consortium Ltd

This is the most comprehensive SNP database. The SNP Consortium

Ltd. is a non-profit foundation organized for the purpose of providing public

genomic data. This database was one of the goals of Genome II, the next

phase of the Human Genome Project (Brower 1998). Its mission was to

develop up to 100,000 SNPs distributed evenly throughout the human genome

and to make the information related to these SNPs available to the public

without intellectual property restrictions. SNP screening is performed at the

three major genomic centers (Washington University at St. Louis, The

Whitehead Institute for Biomedical Research and The Sanger Center), by

using a panel of unrelated, anonymous individuals. The fifth release (April

16

2000) of this database consisted of 296,990 SNPs from all the chromosomes.

In the present thesis, I collected data from the fifth release of this database. An

example of this data is shown in Appendix 3.

ALLELE FREQUENCY DISTRIBUTION

SNPs are generally present in biallelic form, but sometimes they are present in

more than two forms. I investigated the distribution of the frequencies of the most

common alleles in the SNP data obtained from The Genome Database (GDB). I only

considered biallelic SNPs in this data set. The Genome Database has given the allele

frequency for biallelic polymorphisms. I considered those allele frequencies that are

more than 0.5 (More than 0.5 indicates frequent occurrence of that allele).

To check the symmetry and normality of the data, I performed a kurtosis plot

analysis to test if the data are normally distributed and also to verify that the

distribution is not skewed to the left or right of the mean. By using symmetry and

kurtosis measures, the normality of the allele frequency data was assessed. The null

hypothesis of population normality was tested by using the test statistic,

K2 = Z2g1 + Z2

g2

where Z2g1 and Z2

g2 are the parameters for symmetry and kurtosis. K2 has

the same distribution as χ2 and it is the parameter for assessing normality using

symmetry (Zg2) and kurtosis (Zg2) measures. Zg1 is the parameter for testing the

population’s symmetry. But not all symmetrical distributions are normal and to check

the normality, kurtosis measure is used. Here, Zg2 is the population kurtosis

parameter. The calculations of Zg2 and Zg2 are as follows.

17

Zg1 = E ln(F + 1F2 + )

F = 1 C 2A / −÷

E = 1 / Dln

D = C

C = )1 (B2 − - 1

A = 2)-6(n / 3) (n 1) (n b1 ++

where, n = number of SNPs considered and 1b = beta measure of symmetry.

Zg2 =

K92

LK92

1 3−−

L =

4K2

H1

K2

1

−+

−

K =

+++

2J41

J2

J86

J = )3n)(2n(n)5n)(3n(6

)9n)(7n()2n5n(6 2

−−++

+++−

H = G)1n)(1n(

|g|)3n)(2n( 2

−+−−

3(n2 + 27n – 70) (n + 1) (n + 3)

(n – 2) (n + 5) (n + 7) (n + 9) B =

18

G = )5n)(3n()1n(

)3n)(2n(n242 +++

−−

2g is the sample statistic for the measure of kurtosis and for further calculations

of 2g refer (Zar 1999 (b)).

19

SIMULATION STUDY

To study the effect of different ratio of transitions to transversions, I generated

sequence data by using a computer simulation program called TREEVOLVE (Grassly

and Rambaut, unpublished) 1.

Principle of TREEVOLVE

TREEVOLVE simulates DNA sequences based on Kingman’s (1982a,b,c)

coalescent approach (cited in Kingman 2000). The coalescent model simulates the

evolutionary history of a gene backwards in time until it reaches a point of most

recent common ancestor (MRCA) for that gene. This generates a tree for a particular

gene in the population sample. The tracing backwards in time is based on a Markov

chain (Kingman 2000). An example shown in the following figure (figure 4),

MRCA (most recent common ancestor)

Past

Present Generation

1 2 3 4 5

Figure 4: Simulation of population history by TREEVOLVE.

1 http://evolve.zoo.ox.ac.uk

20

According to this figure, if there are n members of a particular generation,

then TREEVOLVE traces their family tree backward through time until it coalesces

(i.e. the lineages find a common ancestor). The number n is reduced by one each time

coalescence occurs so that next time (n-1) lines (members) will be traced back until

they coalesce and so on, until the number of lines is reduced to one (MRCA). While

simulating the tree, the program does not take into account any sequence information

from the present-day generation individuals. The lineages are joined randomly back

in time. Depending on the number of replicates specified in the input parameter file, a

new random tree is generated in each replication. Each tree corresponds to a separate

replicate. After generating each tree, the molecular sequences are simulated down the

genealogy under the substitution model specified by a user.

Substitution model

For each tree, TREEVOLVE generates present-day sequences. The number of

sequences depends on the input parameter file shown in figure 6. The sequences

differ completely between the replicates. After generating each coalescent tree, the

DNA sequences are generated according to the genealogy down the line with time

(forward direction) starting at the most recent common ancestor (MRCA) as shown in

figure 5. TREEVOLVE has an option of using different substitution models based on

the type of analyses and the choice of variables in the sequence generation. Two of

the models are F84 (Felsenstein and Churchill 1996) and HKY85 (Hasegawa et al.

1985) that takes into account the transition to transversion ratio and also the base

frequencies as parameters, but differing in their values. I used the one- and two-

21

parameter models which can be presented as a partial case of F84 model with equal

base frequencies in all simulations. In the case of the one-parameter model which

assumes equal rate of transitions and transversions, I used the transition to

transversion ratio (α/2β) of 0.5. When I used the two-parameter models, the ratio of

α/2β was varied between 1, 2.35, 3 and 5. First of all, for a given set of simulation

parameters, TREEVOLVE generates 200 trees (200 replicates). For each tree, it

generates present-day sequences depending on the selected substitution model. Each

tree corresponds to a separate replicate and these sequences differ completely

between the replicates. While simulating these sequences TREEVOLVE also takes

into account the mutation rate and recombination rate. As a result, some polymorphic

sites within each replicate are produced. An example of an input parameter file for

TREEVOLVE is shown in figure 6, and the complete list of parameters for

TREEVOLVE used in this study is shown in table 1. The simulations were carried

out under no recombination, and also with recombination rates varied between 10-8,

3×10-5 and 10-3 as suggested by other studies (Zollner and Haeseler 2000) keeping all

other parameters constant. I used no population subdivision (m = 0 in figure 6), no

migration, and no exponential growth (e = 0.0 in figure 6) to minimize the effect of

admixture, allelic heterogeneity and environment. The mutation rate was 10-7, which

was constant in all simulations, as suggested by other studies (Zollner and Haeseler

2000). I used a sequence length of 1000bp with 200 sequences (same as sample size)

for each file. The substitution model used was either a one- or two-parameter model

as described earlier. The simulations were performed as shown in table 2. A sample

run of TREEVOLVE is shown in figure 7.

22

MRCA (most recent common ancestor)

1 2 3 4 5

Figure 5: Se quence simulation by TREEVOLVE after the population history has been simulated.

BEGIN TVBLOCK [sequence length] l1000

[sample size] s200 [mutation rate] u0.0000001 [number of replicates] n1 [substitutionmodel] vF84 t0.5 [f0.25,0.25,0.25,0.25] [r0.1667,0.1667,0.1667,0.1667, 0.1667, 0.1667]

[output coalescent times] oCoal.Times [diploid]

[generation time/variance in offspring number] b1.0 *PERIOD 1 [length of period] t100000.0 [population size] n100000 e0.0 [subdivision] d1 m0 [recombination] r0.0 *END

Figure 6: An example of input parameter file for TREEVOLVE.

Past

Present Generation

23

Table 1: Input parameter values for TREEVOLVE used in this study.

Sequence length, l 1000 bp

Sample size, s 200 sequences

Mutation rate, u 0.0000001

Number of replicates, n 200

Substitution model, m One- and two-parameter model

Transition/transversion ratio, t Varies (0.5 to 5.0)

Base frequencies, f Equal (0.25)

Rate heterogeneity None

Ploidy Diploid

Generation time/var. in offspring no., b 1.0

Run time for population dynamic model,t 100000.0

Effective population size, n 100000

Exponential growth rate, e 0.0

Number of demes, d 1

Migration rate, m 0.0

Recombination rate Varies from 0.0 to 10-3

Parameters Value

24

Sequence1 TCAGGAACAACAGCTAATGAGCTTATATTTTCATGACATAACG Sequence2 TCAGGAACAACAGCTAATGAGCTTATATTTTCATGACATAACG Sequence3 TCAGGAACAACAGCTAATGAGCTTATATTTTCATGACATAACG Sequence4 TCAGGAACAACAGCTAATGAGCTTATATTTTCATGACATAACG Sequence5 TCAGGAACAACAGCTAATGAGCTTATATTTTCATGACATAACG Sequence6 TCAGGAACAACAGCTAATGAGCTTATATTTTCATGACATAACG Sequence7 TCAGGAACAACAGCTAATGAGCTTATATTTTCATGACATAACG

Figure 7: A sample output of TREEVOLVE for sample size of 7 with sequence length of 43. There are no SNPs in nucleotide sequences shown because only few initial bases are shown for each sequence for illustration.

Table 2: The table indicating different simulations performed for different

parameters.

Recombination rate Transition

Transversion 0.0 10-8 3×10-5 10-3

0.5 √ √ √ √

1.0 √ √ √ √

2.35 √ √ √ √

3.0 √ √ √ √

5.0 √ √ √ √

25

LINKAGE DISEQUILIBRIUM ANALYSIS OF SIMULATED

DATA FROM TREEVOLVE

Different types of software are used in order to compute values of LD (linkage

disequilibrium) using the sequence data. In the present study, program Arlequin1

(Schneider et al. 1997) was used for calculating the LD in the data simulated by

TREEVOLVE. It is versatile software available for analysis of genetic data of various

different types, including RFLPs, DNA sequence, and microsatellite data. It allows

one to perform a number of statistical tests using population data, including linkage

disequilibrium analysis. Arlequin has a graphical interface that is user-friendly. Its

other advantage is that one can run a number of input files simultaneously with the

same or different parameter lists by creating a batch file, thus reducing the analysis

time of the user. An example of an input file of Arlequin used in this study is shown

in figure 8.

During analysis by Arlequin, there were 200 input files (replicates from

TREEVOLVE) for each set of parameters. All these input files were analyzed

together by creating a batch file (figure 9).

1 http://anthro.unige.ch/arlequin/

26

[Profile]

Title="SNPAnalysis Ratio5 r0.0" NbSamples=1 GenotypicData=0 DataType=DNA LocusSeparator=NONE MissingData='?'

[Data] [[Samples]] SampleName="Replicate1" SampleSize=10 SampleData= {

sequence1 1 ACAACAGCTAATGAGCTTATATTTTCATGACATAACGGGAAC sequence2 1 ACAACAGCTAATGAGCTTATATTTTCATGACATAACGGGAAC sequence3 1 ACAACAGCTAATGAGCTTATATTTTCATGACATAACGGGAAC sequence4 1 ACAACAGCTAATGAGCTTATATTTTCATGACATAACGGGAAC sequence5 1 ACAACAGCTAATGAGCTTATATTTTCATGACATAACGGGAAC } Figure 8: Sequence.arp – An input file for Arlequin for the calculation of LD. The field under SampleData is the output of treevolve with transition to transversion ratio of 0.5 and recombination rate of 0.0 (Only first five sequences are shown with only forty nucleotides in each sequence. The data simulated by TREEVOLVE had 200 such sequences with 1000 nucleotides in each sequence for each set of parameters).

27

replicate1.arp replicate 2.arp replicate 3.arp replicate 4.arp replicate 5.arp replicate 6.arp replicate 7.arp replicate 8.arp replicate 9.arp replicate 10.arp

Figure 9: Batch.arb – An example of a Batch file for Arlequin. During analysis, 100 to 200 replicates were analyzed in each batch file.

Arlequin calculates LD by doing pair-wise comparisons between different

SNPs within each replicate. The LD is calculated by using the formula,

Dij = ρij - ρi ρj,

where ρij is the frequency of the haplotype having allele i at the first locus and allele j

at the second locus, and ρi and ρj are the frequencies of alleles i and j in the replicate

respectively.

After obtaining the linkage disequilibrium (D) values for all the input

parameter files, I calculated the mean D and standard deviation (SD) values for all the

combined 200 replicates for each parameter file (i.e. for different transition to

transversion ratios at different recombination rates).

28

RESULTS

SUMMARY OF DATA COLLECTED

In order to test whether the ratio of transitions to transversions is 0.5 (α/2β) as

predicted, I collected SNP data from three databases. The summary of the SNP data

collected is given in table 3. The results indicate tha t the transition to transversion

ratio in the real data is at least four times higher than the expected ratio of 0.5.

Table 3: Data Summary from all the three databases

Dataset Number of SNPs α/2β The Whitehead Institute of

Genome Research 146 2.35

The Genome Database 36 2.003 The SNP Consortium 5102 2.086

1) The Whitehead Institute of Genome research

The SNP data is shown in appendix 1. The distribution of SNPs in this data is

shown below in figure 10. The observed ratio of α/2β is 2.35, which is approximately

five times higher than the expected ratio of 0.5. The total number of biallelic SNPs in

this dataset was 146.

29

Proportion of SNPs:

Proportion

A/G, G/A: 25+24 = 49 0.336 A/C, C/A: 10+4 = 14 0.096 A/T, T/A: 3+5 = 08 0.055 G/C, C/G: 6+7 = 13 0.089 G/T, T/G: 5+3 = 08 0.055 C/T, T/C: 24+30 = 54 0.370

------------ Total = 146

Proportion of transitions and transversions: Proportion Transitions- A/G + G/A + C/T + T/C = 103 0.705 Transversions- A/T + T/A + A/C + C/A + G/T + T/G + G/C + C/G = 43 0.300

Figure 10: SNP distribution in the chromosome 6 data from The Whitehead Institute of Genome Research.

2) The Genome database

The SNP data from this database are shown in appendix 2. The distribution of

SNPs in this database is shown in figure 11. The observed ratio of α/2β is 2.003,

which is four times higher than the expected ratio of 0.5. There were 36 biallelic

SNPs in this dataset.

30

Proportion of SNPs:

Proportion

A/G, G/A: 06+07 = 13 0.361 A/C, C/A: 01+01 = 02 0.055 A/T, T/A: 01+01 = 02 0.055 G/C, C/G: 04+01 = 05 0.139 G/T, T/G: 02+01 = 03 0.083 C/T, T/C: 10+01 = 11 0.305

----------- Total = 36

Proportion of transitions and transversions:

Proportion Transitions - A/G + G/A + C/T + T/C = 24 0.667 Transversions - A/T + T/A + A/C + C/A + G/T + T/G + G/C + C/G = 12 0.333 Figure 11: SNP distribution in the chromosome 6 data from The Genome Database.

3) The SNP Consortium Database

An example of the SNP data from the fourth release of the SNP Consortium

Database is shown in appendix 3. The distribution of SNPs in this dataset is given in

figure 12. The total number of biallelic SNPs in this data set was 5102. The observed

ratio of α/2β is 2.09 which is four times higher than the expected ratio of 0.5. The

total number of biallelic SNPs in this data set is 5102. This number is much higher

than that for the first two data sets.

31

Proportion of SNPs: Proportion

A/G, G/A: 1750 0.343 A/C, C/A: 425 0.083 A/T, T/A: 342 0.067 G/C, C/G: 432 0.085 G/T, T/G: 454 0.089 C/T, T/C: 1699 0.333

----------- Total = 5102

Proportion of transitions and transversions:

Proportion Transitions - A/G + G/A + C/T + T/C = 3449 0.676 Transversions - A/T + T/A + A/C + C/A + G/T + T/G + G/C + C/G = 1653 0.324 Figure 12: SNP distribution in the chromosome 6 data from The SNP Consortium.

The results showing the number of SNPs are summarized in table 3. This table

indicates that transitions are much more common than transversions, and therefore a

two-parameter model may better describe the SNP properties than the one-parameter

model. The overall results from figures 10, 11, and 12 are summarized in table 4.

Table 4: Summarized results from figures 10, 11, and 12.

Database Transitions Transversions Transitions/transversions

1 103 43 0.705/0.300 = 2.35

2 24 12 0.667/0.333 = 2.003

3 3449 1653 0.676/0.324 = 2.086

32

ALLELE FREQUENCY DISTRIBUTION

The main aim was to understand the distribution of SNP variants in human

population. This knowledge may be helpful in the future when generating simulated

SNP data.

To see the distribution of the allele frequencies in the human population of the

SNPs, I analyzed the allele frequency data collected for the SNPs from the Genome

Database. I selected those alleles which were common (i.e. they have a frequency of

more than 0.5). The distribution of allele frequencies is shown in figure 13. This

distribution suggests that the allele frequency of most common alleles in SNP data

follows a normal distribution with mean allele frequency of 0.7520 and standard

deviation of 0.1272.

From the Kurtosis plot analysis performed to see the symmetry and normality

of the allele frequency data,

K2 = 2.016915 (0.25 < P < 0.50) (From the Chi-squared distribution)

table with two degrees of freedom)

Thus, the result of K2 indicates that the data are normally distributed. These

results will help in the future for simulations of SNP data while considering the allele

frequencies.

33

Figure 13: Allele frequency distribution for all common alleles of SNPs from human chromosome 6.

Allele frequency distribution

02468

1012

0.52

0.58

0.64 0.7 0.7

60.8

20.8

80.9

4M

ore

Allele frequency

Freq

uenc

y

34

EFFECT OF THE RATIO OF TRANSITIONS TO

TRANSVERSIONS ON LINKAGE DISEQUILIBRIUM

In order to study the effect of different ratios of transition to transversion on

linkage disequilibrium, which is a frequently used parameter in association analysis, a

simulation study was performed. After running the simulations for each set of

parameters and analyzing the simulated data by Arlequin, I calculated the mean

linkage disequilibrium values for each set of parameters. The results of this analysis

are shown in the table 5.

Table 5: Mean linkage disequilibrium values ±± standard deviation for different values of the transition to transversion ratios at different recombination rates.

Recombination rate Transition/transversion

(αα /2ββ) 0.0 10-8 3××10-5

0.5 0.0178 ±0.0393 0.0221 ±0.0454 0.0001 ±0.0006

1.0 0.0214 ±0.0420 0.0236 ±0.0495 0.00014 ±0.0007

2.35 0.0211 ±0.0426 0.0175 ±0.0423 0.00011 ±0.0006

3.0 0.0193 ±0.0394 0.0205±0.0475 0.00008 ±0.0006

5.0 0.0183 ±0.0386 0.0158 ±0.0391 0.00012 ±0.0007

The relationships between different transition to transversion ratios and mean

LD at three recombination rates of 0.0, 10-8 and 3×10-5 are shown in figures 14, 15

and 16, respectively, with error bars indicating standard deviation (SD). No variable

loci were observed for data simulated under the highest recombination rate (10-3). I

also observed a variation in the number of polymorphic loci for different

35

recombination rates, with the number of SNP loci decreasing as the recombination

rate increased.

36

Figure 14: Plot of Mean LD Vs. αα /2ββ at recombination rate (r) equal to 0.0.

Error bars indicate the standard deviations (SD)

Figure 15: Plot of Mean LD Vs. αα /2ββ at recombination rate (r) equal to 10-8.

Error bars indicate the standard deviations (SD)

37

Figure 16: Plot of Mean LD Vs. αα /2ββ at recombination rate (r) equal to 3××10-5

Error bars indicate the standard deviations (SD)

38

DISCUSSION

According to the one-parameter model, the expected ratio of transitions to

transversions is 0.5 (since the transition rate is expected to be equal to the rate of the

transversion). To see the actual proportion of transitions to transversions in the

general human population, I collected data from publicly available databases. The

ratio observed from the SNP data from the public databases showed that transitions

are more frequent than transversions (transitions are approximately four times higher

than transversions). Therefore, the two-parameter model (where transition and

transversion rates differ) may better approximate the SNP behavior. My analyses

showed that the proportion of transitions to transversions was very different (70% to

30%, 66% to 33%, and 68% to 32% in the Whitehead Institute of Genome Research,

the Genome Database and the SNP Consortium databases respectively) than that

expected (33% to 66%) under the scenario of no differences in mutation rates

between different nucleotide changes.

To study the distribution of the frequencies of the most common alleles in the

SNP data obtained from The Genome Database (GDB), I considered all biallelic

SNPs in this data set. The allele frequency distribution of the most common alleles

turned out to be normal with mean allele frequency of 0.7520 and standard deviation

of 0.1272. After performing the Kurtosis plot analysis on this allele frequency data, it

showed that the distribution of allele frequency is symmetric as well as normal (K2 =

2.016915, corresponding to 0.25 < P < 0.50). This analysis will help in the future for

simulating SNP behavior, because allele frequencies play an important role in

detecting strong association between a marker SNP and a susceptibility loci. If

39

marker allele frequencies are substantially different from susceptibility allele

frequencies, then one needs a large sample size or a large number of markers or both

to have a strong association (McCarthy and Hilfiker 2000). Therefore, the allele

frequency distribution of common SNPs would help for future simulation studies

involving SNP behavior in application to association studies involving allele

frequencies.

I analyzed the effect of different transition to transversion ratios (α/2β) on

linkage disequilibrium. Results from the linkage disequilibrium analysis of simulated

population showed that the linkage disequilibrium remained approximately same for

different transition to transversion ratios for the parameters used in the simulations.

The results obtained were for a sequence lengths of 1000 bp. As the recombination

rate increased, the mean LD value decreased. The increase in recombination rate also

resulted in reduction of number of polymorphic loci, thus reducing the strength of

LD. The reduction in number of loci could be related to the small sequence length

(1000 bp) in simulations and high recombination rate. An average extent of useful

levels of LD in the general human population is approximately 3kb as shown by

Kruglyak (1999). It is possible that the α/2β ratio might have an effect on LD when

longer sequence length is used. Longer sequences can be investigated in future

studies. One can also analyze the transitions and transversions separately from the

SNP data obtained from the simulation studies. In the present study, the LD was

evaluated between the polymorphic loci without considering whether each SNP was a

transition or transversion.

40

No polymorphic sites were observed in the sequences simulated by

TREEVOLVE for recombination rate of 10-3. This result was observed with short

sequence length and high recombination rate, which reduces the strength of LD

considerably, because the chances of a polymorphism being fixed in a population are

considerably less at a higher recombination rate while simulating the sequences. This

phenomenon is because of the way the TREEVOLVE simulates the data. The

program tries to simulate a polymorphism in the population while generating the

sequences and not while simulating the population tree. When the recombination rate

is much higher than the mutation rate, TREEVOLVE simulates a population that does

not have polymorphism. These results support those previously obtained by Kruglyak

(1999), in an independent analysis, in which he found no LD at recombination rate of

3×10-4 or higher (corresponding to physical distance of approximately 30kb). At such

a high rate of recombination, there is higher separation between polymorphic loci.

Such a high separation is probably unable to be accommodated in a sequence length

of 1000 bp, resulting in the absence of any polymorphic loci in the population.

Part II

Application of SNPs in fine gene mapping

41

INTRODUCTION

As described in the earlier part of this thesis, association analysis has been of

considerable importance in various fields ranging from population genetics,

pharmacogenomics, population genetic epidemiology and toxicology. An association

is said to exist between two phenotypes (of which one is resulting into a disease) if

they occur in the same individual more often than expected by chance. To investigate

whether the two phenotypes are associated or not, one collects the two groups of

individuals, one with the affected individuals and one for controls (unaffected

individuals). Then by counting the proportion of individuals having disease and the

other phenotype and individuals with disease and not having the other phenotype, one

can perform a standard 2×2 chi-square allelic association analysis or a 3×2 chi-square

genotypic association analysis to examine the significance of the association.

A combination of alleles at a specific gene characterizes a genotype of an

individual. Alleles and genotypes play a very important role in association analysis. If

a gene is a disease pre-disposing gene, it is possible that only one of the alleles of that

gene is actually responsible for the disease predisposition. Therefore, an allelic

association was a common way of fine gene mapping until recently, although allelic

heterogeneity may complicate this method (Terwilliger and Weiss 1998). Thus,

genotypic association analysis can be used in place of allelic association. In allelic

association, alleles are used to identify an association with disease. In contrast, in

genotypic analysis one looks at association of a disease with genotypes (similar to

genotypic linkage disequilibrium described by Weir 1996). In genotypic association

42

analysis one studies differences in genotype frequencies in healthy and affected

individuals to find if any genotype is associated with the disease.

Objective

In this part, I describe an approach for fine gene mapping using SNPs from a

general population by using genotype-specific disequilibrium analysis. The approach

is different from the allelic disequilibrium analysis, because it considers the frequency

of a genotype and not of individual alleles.

Description of the simulated data

A committee of Genetic Analysis Workshop 12 (GAW12) organized by

Southwest Foundation for Biomedical Research, San Antonio, Texas, USA, simulated

the data for GAW12 2000. The data used in this thesis were simulated for a large

general population. The disease prevalence in the population was about 25%. The

data were provided for 23 extended pedigrees with 1497 total individuals (1000

living) in the population. The disease was more prevalent in females than in males.

Five quantitative risk factors (Q1 to Q5) and two environmental factors (E1 and E2)

were also associated with the disease. Seven major genes (MG1 to MG7) influence

these five quantitative risk factors as shown in the figure 17. A major gene is a gene

related to the disease risk. The overall summary of the generating model is shown in

figure 17. This model was not known during the analysis. In the data available to me,

each living individual had information on affectation status, age at last exam, age at

onset of disease if affected, five quantitative risk factors and two environmental

43

factors as well as genotypic data for SNPs present in 7 candidate genes. These genes

were named from 1 through 7. These seven candidate genes were the genes, which

were potential candidates that might affect the disease. The data for these seven

candidate genes were provided by the GAW12 for analysis. These candidate genes

were present in the major genes. The goal of the study described in this thesis was to

test whether any of these candidate genes were contributing to the disease, and to

identify any functional SNPs that could be related to the disease risk. The data were

provided for 50 such replicates (50000 living individuals). There were 165 founders

in each replicate. There was a total of 9515 original SNPs in the population of 50,000

individuals. The organizers also provided us with the information about which was

the best replicate in the data. The best replicate was replicate 42 that had the data with

contributions from all the factors discussed above. It represented the best sample

population among all 50 replicates, with the mean simulated parameters being the

closest to the parameters originally used I simulations.

44

Figure 17: The phenotypic model of the data simulated by GAW12. Boxes indicate items provided in the GAW12 data set, including quantitative traits, affectation status, age at onset, environmental factors (E1 and E2) and household membership (HH). Circles indicate genetic factors including seven major genes and a mitochondrial (Mito) component.

MG5 MG1

MG6

45

METHODS

Individuals used in the study

The study was performed on pedigree founders as well as all the living

pedigree members. A founder is an individual in a family tree who does not have any

living ancestors and is not related to anyone in his or her generation (Figure 18).

Figure 18: Founders in a pedigree.

I considered founders initially, because they were unrelated. Such an approach

minimizes the correlations among family members in the SNP association analysis.

Several data sets were used in this study.

1) 8250 pedigree founders from all 50 replicates studying all candidate genes

1 to 7.

FOUNDERS

NONFOUNDERS

DEAD

DEAD

46

2) 8250 pedigree founders from all 50 replicates, analyzing only for genes 1

and 2 separately.

3) 165 pedigree founders from the best replicate 42, using only genes 1 and 2

separately.

4) 1000 living individuals from the best replicate 42 using only genes 1 and 2

separately.

Genotypic and phenotypic data

I considered the genotypic data for biallelic SNPs from 7 candidate genes for

each individual along with his or her affectation status. The genotypic data were for

715 candidate SNPs selected after the data reduction. The data for each SNP genotype

were provided by the GAW12 in binary format, i.e. 11, 12, or 22 instead of the

nucleotides. Therefore, the transition or transversion nature of the polymorphisms

was not taken into account.

Data reduction

The genotypic data set was very large. Considering all the 50 replicates, there

were 50000 individuals. To minimize the dimensionality of the data set, only those

SNPs were considered that were present in the pedigree founders in each of the 50

replicates by using a program called DATACONVERT (SaoPedro, unpublished). As

a result, 715 SNPs were obtained from the total of 9515 SNPs after the data

reduction. I considered the SNPs from the both coding and non-coding regions in the

analysis.

47

Sorting technique

The first step in the analysis was to sort the data. I considered affection status

as the variable of interest. I used program Scansort (Ward 2000, unpublished), which

sorts the data according to the sum of the absolute differences between frequencies of

healthy and affected subjects with the three SNP genotypes (11, 12, or 22, where 1 is

a wild type, or ancestral, allele and 2 is a mutated allele). The output of the program

is an eleven-dimensional data set, which contains one record for each SNP position in

the following order:

a. SNP index (original order)

b. Frequency of healthy individuals with genotype 11 (f11h)

c. Frequency of healthy individuals with genotype 12 (f12h)

d. Frequency of healthy individuals with genotype 22 (f22h)

e. Frequency of affected (sick) individuals with genotype 11 (f11s)

f. Frequency of affected (sick) individuals with genotype 12 (f12s)

g. Frequency of affected (sick) individuals with genotype 22 (f22s)

h. Sum of all the absolute differences (d)

i. Absolute difference between c and f

j. Absolute difference between d and g

k. Absolute difference between e and h

Table 6 represents this information in tabular form.

48

Table 6: Calculation of maximum difference by Scansort program.

Individual Genotype

11 12 22

Healthy f11h f12h f22h

Sick (Affected) f11s f12s f22s

Scansort calculates d, the sum of the absolute differences, by using the

following formula.

d = (| f11h – f11s| + |f12h-f12s| + |f22h-f22s|) / Number of individuals in the data set

Scansort sorts the SNP data according to the d value. Therefore, Scansort

orders the SNP data in a useful way where SNPs that have very different proportions

in healthy and affected individuals are present at the top of the list. As a result, the

output of the Scansort was used for the conventional 3×2 chi-square analysis.

Comparative statistical analysis of genotype frequencies

I used the chi-square statistic to find out the significant SNPs associated with

the disease. The Bonferroni correction was applied to correct the significance levels

for multiple testing. Statistical analysis was performed using Microsoft Excel.

Chi-square statistic is calculated by using the following formula,

(observed frequency – expected frequency) 2

expected frequency χ2

=

49

The chi-square analysis for individual SNP position was performed by

considering a 3×2 contingency table (Table 7).

Table 7: Chi-square analysis for an individual SNP position.

Genotype Individual

11 12 22

Healthy f11h f12h f22h R1 = f11h + f12h + f22h

Sick (Affected) f11s f12s f22s R2 = f11s + f12s + f22s

C1 = f11h + f11s C2 = f12h + f12s C3 = f22h + f22s A = C1+C2+C3 = R1+R2

Let A = C1+C2+C3 = R1+R2 (total individuals in a population), then

Expected frequency for f11h, f11he = C1×R1/A

The chi-square distribution is calculated by using appropriate degrees of

freedom. The degree of freedom is calculated by using following formula (Zar 1999),

Degrees of freedom = (columns – 1) × (rows – 1)

The Bonferroni correction was used to adjust the significance level, as there

are a large number of tests to be performed because of the high number of loci (Weir

1996). α was calculated by using following formula.

α = 1- (1 - α’)1/L

where α’ = 0.05, and L is the number of SNPs analyzed.

(Observed frequency–expected frequency)2

(expected frequency) = ∑ χ2

50

RESULTS

To find the significant SNPs associated with the disease I performed chi-

square analysis initially on a data set containing the founders from all the 50

replicates. Each of the 715 SNPs in genes 1 through 7 were analyzed for the

differences in their genotype frequencies between healthy and affected subjects for

the data set with all the pedigree founders from the 50 replicates. These SNPs were

then sorted by the Scansort program for the sum of their absolute differences between

their genotype frequencies. The top twenty SNPs from this analysis are shown in

table 8.

Table 8: Scansort output for data set 1 (8250 pedigree founders) for top 20 SNPs

sorted by the value of d. All the 20 SNPs are from gene 1.

SNP ID 11h 12h 22h 11s 12s 22s d (diff)557 93 1657 4103 368 1281 748 0.77790376 4113 1647 93 754 1278 365 0.776313

2619 4219 1547 87 864 1235 298 0.7207531553 4236 1536 81 871 1229 297 0.7207213573 4197 1565 91 863 1236 298 0.714073835 4189 1572 92 860 1238 299 0.7138393853 92 1575 4186 302 1236 859 0.7136483742 92 1576 4185 302 1236 859 0.7133073456 4238 1530 85 889 1224 284 0.7063865757 4122 1620 111 842 1240 315 0.7059647281 4081 1654 118 838 1237 322 0.6952912942 139 1738 3976 330 1264 803 0.6886152923 137 1766 3950 336 1258 803 0.679731

11180 4007 1710 136 837 1229 331 0.6708391478 172 1845 3836 356 1257 784 0.65663189 268 2154 3431 468 1303 626 0.650071596 3433 2151 269 627 1309 461 0.64992

4471 200 1913 3740 360 1272 765 0.6396794752 203 1922 3728 361 1273 763 0.6372483534 262 2079 3512 396 1274 727 0.593477

51

After having obtained the ordered SNP data for 715 SNPs from all the genes

by Scansort, I performed chi-square analysis on the frequency data obtained from the

Scansort output. The results for the most significant SNPs according to the p-values

in genes 1, 2, and 6 are shown in tables 8, 9, and 10, respectively. I did not found any

significant SNP in any other genes except 1,2, and 6. The resulting Bonferroni

correction for each SNP was 7.17363×10-5 when the initial significance level was set

to 0.05.

Table 9: The 20 most significant SNPs in gene 1. The data set analyzed was the

8250 pedigree founders from all the 50 replicates.

SNP ID 11h 12h 22h 11s 12s 22s Chi Square p-value557 93 1657 4103 368 1281 748 1315.64227 2.0507*10-286

76 4113 1647 93 754 1278 365 1308.13783 8.7394*10-285

1553 4236 1536 81 871 1229 297 1124.2512 7.4465*10-245

2619 4219 1547 87 864 1235 298 1112.56292 2.5706*10-242

3853 92 1575 4186 302 1236 859 1090.46042 1.62*10-237

3742 92 1576 4185 302 1236 859 1089.66349 2.4131*10-237

3573 4197 1565 91 863 1236 298 1088.87717 3.5754*10-237

3835 4189 1572 92 860 1238 299 1087.27638 7.9603*10-237

3456 4238 1530 85 889 1224 284 1068.70713 8.5742*10-233

5757 4122 1620 111 842 1240 315 1052.39334 2.9901*10-229

7281 4081 1654 118 838 1237 322 1024.87688 2.8236*10-223

2942 139 1738 3976 330 1264 803 984.310527 1.8183*10-214

2923 137 1766 3950 336 1258 803 976.29031 1.0028*10-212

11180 4007 1710 136 837 1229 331 954.378676 5.7451*10-208

189 268 2154 3431 468 1303 626 916.243316 1.0972*10-199

596 3433 2151 269 627 1309 461 905.951689 1.884*10-197

1478 172 1845 3836 356 1257 784 902.347863 1.1419*10-196

4471 200 1913 3740 360 1272 765 838.786422 7.2417*10-183

4752 203 1922 3728 361 1273 763 831.839741 2.335*10-181

12185 3543 2059 251 747 1237 413 750.82136 9.1456*10-164

52

Table 10: The 15 most significant SNPs in gene 2. The data set analyzed was the

8250 pedigree founders from all the 50 replicates.

Table 11: The 12 most significant SNPs in gene 6. The data set analyzed was the

8250 pedigree founders from all the 50 replicates.

SNP ID 11h 12h 22h 11s 12s 22s Chi Square p-value4894 680 2643 2530 170 873 1354 127.7856 1.7853*10-28

4977 3605 2003 245 1782 565 50 125.28594 6.2302*10-28

4766 672 2634 2547 172 873 1352 120.04499 8.5617*10-27

3185 614 2594 2645 158 848 1391 117.5252 3.0180*10-26

861 569 2556 2728 145 828 1424 116.38561 5.3356*10-26

1495 567 2555 2731 144 829 1424 115.68125 7.5881*10-26

715 2730 2554 569 1423 830 144 115.47719 8.4032*10-26

4030 627 2597 2629 158 865 1374 112.08614 4.5793*10-25

4538 633 2604 2616 167 859 1371 111.29881 6.7884*10-25

5961 3826 1834 193 1844 518 35 110.24763 1.1482*10-24

2540 570 2539 2744 152 856 1389 88.29132 6.7264*10-20

3155 2914 2454 485 1454 814 129 84.504551 4.4675*10-19

5219 957 2869 2027 281 1059 1057 73.380148 1.1633*10-16

2805 738 2684 2431 216 948 1233 71.968459 2.356*10-16

5499 978 2877 1998 287 1073 1037 70.221367 5.6444*10-16

SNP ID 11h 12h 22h 11s 12s 22s Chi Square p-value6805 4948 873 32 1725 625 47 185.364929 5.6042*10-41

7332 4948 873 32 1725 625 47 185.364929 5.6042*10-41

8067 4951 870 32 1727 623 47 184.886869 7.1174*10-41

5782 4947 874 32 1725 625 47 184.845233 7.26721E-41

7073 4946 874 33 1724 626 47 184.492956 8.6668*10-41

5007 4942 879 32 1724 626 47 183.275133 1.5933*10-40

8226 35 893 4925 47 633 1717 179.169005 1.2414*10-39

1987 202 1793 3858 146 879 1372 67.4235199 2.2864*10-15

1748 3855 1797 201 1372 879 146 67.1851486 2.5759*10-15

993 3854 1796 203 1373 877 147 66.4377271 3.7430*10-15

11782 513 2420 2920 155 909 1333 26.8361224 1.4880*10-06

13869 514 2415 2924 158 911 1328 24.2681663 5.3732*10-06

53

The range of significant SNPs for each gene is reported in table 12. According

to the significance level of 7.17363×10-5 (Bonferroni cut-off correction), only 172

SNPs out of 715 SNPs were significant, and thus could be associated with the disease.

Table 12. Number of significant sequence polymorphisms, the range of their

significance values, and the total number of polymorphisms, by gene,

determined from the 8250 pedigree founders from all the 50 replicates.

Gene # of significant SNPs Range of significant p-values Total # of SNPs

1 107 2.0507x10-286 to 6.21184x10-5 157

2 52 1.78529x10-28 to 4.29522x10-5 90

6 13 5.60425x10-41 to 5.37322x10-6 34

I considered only founders initially as they were responsible to transmit the

SNPs if any in the next generation. Hence I analyzed the SNP data considering all the

founders from the 50 replicates in the hope to find all the SNPs significant with

respect to the disease causing polymorphisms. Analysis of all the pedigree founders

showed that most of the SNPs in genes 1 and 2 might be associated with the disease.

Therefore, to narrow down the SNPs of most interest, I analyzed 8250 pedigree

founders for SNPs only in genes 1 and 2 separately following the same procedure as

in 1. The results of this analysis are shown in table 13.

54

Table 13: Number of significant SNPs in genes 1 and 2 analyzed separately for

8250 pedigree founders. (cut-off p-values from Bonferroni correction are

0.000327 for gene 1 and 0.000570 for gene 2)

Gene # of significant SNPs Range of p-values Total # of SNPs

1 114 2.0507x10-286 to 0.000287 157

2 61 1.78529x10-28 to 0.000398 90

I also analyzed the SNP data for genes 1 and 2 separately for the founders in

best replicate 42 in an attempt to identify the most significant SNPs associated with

the disease in the best replicate and then to compare it with the significant SNPs

obtained after analyzing 8250 pedigree founders. These results are shown in table 14.

Table 14: Number of significant SNPs in genes 1 and 2 analyzed separately for

165 pedigree founders in best replicate 42. (cut-off p-values from Bonferroni

correction are 0.000327 for gene 1 and 0.000570 for gene 2)

I also analyzed the SNP data from all individuals in replicate 42 for genes 1

and 2 separately and then compare it with the results obtained for the same analysis

performed with 8250 pedigree founders and 165 founders from replicate 42

Gene # of significant SNPs Range of p-values Total # of SNPs

1 35 9.54x10-8 to 0.000325 157

2 0 -- 90

55

separately to see the effect of sample size on the number of significant SNPs. The

results are shown in table 15.

Table 15: Number of significant SNPs in genes 1 and 2 analyzed separately for

1000 individuals in best replicate 42. (cut-off p-values from Bonferroni

correction are 0.000327 for gene 1 and 0.000570 for gene 2)

Gene # of significant SNPs Range of p-values Total # of SNPs

1 55 2.18x10-35 to 0.000125 157

2 1 6.01x10-6 90

56

DISCUSSION

The analysis was performed to find the most significant SNPs associated with

the disease in the simulated GAW12 data set. I used a genotype disequilibrium

approach in which I considered the difference in genotype frequencies of the SNPs.

According to the results obtained, I found possible mutations in genes 1, 2, and 6 that

were associated with the disease. This analysis was performed without the GAW 12

answers, which were made available to us after the analysis was completed.

The chi-square analysis on data set of 8250 pedigree founders identified 107

significant SNPs in gene 1 (out of a total of 157 SNPs in that gene), 52 SNPs in gene

2 (out of a total of 90 SNPs in gene 2) and 13 SNPs in gene 6 (out of a total of 36

SNPs in gene 6). Because genes 1, 2, and 6 contain multiple polymorphisms which

are associated with the disease, it is difficult to identify causative mutations for the

disease. Hence, one hypothesis that I pursued was that the sequence polymorphisms

showing the lowest p-values were the most likely candidates for affecting the disease

state. In this regard, I identified SNPs at nucleotide position 557 in gene 1 (lowest p-

value of 2.05x10-286), nucleotide positions 6805 and 7332 in gene 6 (p-value =

5.60x10-41) and nucleotide position 4894 in gene 2 (p = 1.79 x10-28) that were most

likely associated with the disease. Many of the SNPs located near these mutations

also have very low p-values, most likely because of the linkage disequilibrium

between SNP variants. For example, in gene 1 the SNP at position 557 has the lowest

p-value of 2.05 x10-286. This is the sixth polymorphism occurring in this gene,

counting from the start of the gene. The second polymorphism from the start of this

gene, at position 76, has the next lowest p-value of 8.74 x10-285. The third

57

polymorphism from the start of the gene, at nucleotide position 189, has a p-value of

1.10 x10-199, and the seventh polymorphism, at position 596, has p-value = 1.88 x10-

197. The fourth, eighth, ninth and tenth polymorphisms at nuc leotide positions 286,

610, 730 and 885 respectively, also show significant association with the disease

(with p-value<0.000327). This observation suggests a likely functional role of the 5'

end of gene 1 in the disease.

There was generally a good correlation between the SNPs identified by the

chi-square test and the polymorphisms that had the top scores identified by program

Scansort. The majority of SNPs that showed significant association with the disease

by the chi-square test were located in the top portion of the arrays generated by

Scansort. The most significant SNPs from genes 1, 2, and 6 (sequence positions 557

in gene 1, 4894 in gene 2 and 6805 and 7332 in gene 6, respectively) were ranked as

number 1, 61, 47 and 48 when sorted by their chi-square values in data set of 8250

pedigree founders. After sorting by Scansort in the same data set for the sum of the

absolute differences in genotype frequencies in healthy and affected individuals, their

positions were 1, 48, 56 and 57, respectively, in the sorted data set. When the 172 top

SNP positions that were significant in data set containing all pedigree founders from

all the 50 replicates according to the chi-square method were compared to the top 172

SNP positions identified by the Scansort, all but 24 SNPs were found to be shared

between the two lists, with the proportion of shared SNPs equal to 87%.

The sensitivity of our tests seemed to increase with the increase of the number

of individuals in the data set. This was the case whether relatives or non-relatives

were added to the data set. I found 25 SNPs in gene 1 which were significant in data

58

set 1 with founders from all 50 replicates, but not significant in either data set 3 or 4.

Analysis of the data set containing all the 715 SNPs with all the founders from 50

replicates revealed 107 significant SNPs in gene 1, while that number was 114 in data

set containing the SNPs for gene 1 only with all the founders from 50 replicates, 35 in

data set containing just founders from replicate 42, and 55 in data set containing all

the individuals in replicate 42. In case of gene 2, I identified 52 significant SNPs in

data set containing the 715 SNPs with all the founders from the 50 replicates, 61

significant SNPs in data set containing all the founders from 50 replicates for gene 2

only, 0 in data set containing just founders from replicate 42 and 1 in data set which

contained all the individuals in replicate 42. The p-values were also smaller in the

larger data sets. Therefore, it seems beneficial to add individuals to the data set even

if they are related to one another.

The answers to this problem provided by the GAW12 organizers (GAW12

Abstracts) were very close to the answers obtained in this analysis. The SNPs

associated with the disease were in genes 1, 2, and 6 according to the GAW12

answers (Figure 17) where candidate gene 1 corresponds to the major gene 6,

candidate gene 2 corresponds to the major gene 5 and candidate gene 6 corresponds

to major gene 1. My analysis also showed association of these three genes with the

disease. The exact positions of the SNPs in these genes according to the answers were

at nucleotide position 557 in gene 1, and position 5782 in gene 6. In gene 2, there

were number of multi-allelic functional variants; in the regulatory elements or in the

first or second bp of a codon; leading to amino acid substitutions in gene 2 and were

associated with the disease. The effect of these genes is shown in figure 17. The

59

analysis presented here also showed the SNP at 557 position in gene 1 as the most

likely candidate; however, the SNPs in gene 6 identified at positions 6805 and 7332

were not correct when compared to the GAW12 answers. Despite discrepancies

related to individual SNPs, the analysis presented here correctly identified the

importance of all the three genes1, 2, and 6 in disease, and these results show

importance of genotypic linkage disequilibrium methods in fine gene mapping.

60

BIBLIOGRAPHY

1) Balendran, N. et al. Characterization of the major susceptibility region for

psoriasis at chromosome 6p21.3. J. Invest. Derm. 113: 322-328 (1999).

2) Brower, V. Genome II: The next frontier. Nature Biotechnology 16(11): 1004-

1007 (1998).

3) Brown, K. The human genome business today. Scientific American: 50-55

(July 2000).

4) Cao, Q., Martinez, M., Zhang, J., Sanders, A. R., Badner, J. A., Cravchik, A.,

Markey, C. J., Beshah, E., Guroff, J. J., Maxwell, M. E., Kazuba, D. M.,

Whiten, R., Goldin, L. R., Gershon, E. S., Gejman, P. V. Suggestive evidence

for a schizophrenia susceptibility locus on chromosome 6q and a confirmation

in an independent series of pedigrees. Genomics 43: 1-8 (1997).

5) Collins, F. S., Guyer, M. S., and Chakravarti, A. Variations on a theme:

Cataloging human DNA sequence variations. Science 278: 1580-1581 (1997).

6) Davies, J. L. A genome-wide search for human type 1 diabetes susceptibility

genes. Nature 371: 130-136 (1994).

7) Elston, R. Linkage and association. Genetic Epidemiology 15: 565-576

(1998).

8) Felsenstein, J. and Churchill, G. A. A hidden markov model approach to

variation among sites in rate of evolution. Mol. Biol. Evol. 13: 93-104 (1996).

61

9) Grassly and Rambaut, TREEVOLVE v. 1.3 (http://evolve.zoo.ox.ac.uk),

unpublished.

10) Hasegawa, M., Kishino, H. and Yano, T. Dating the human-ape splitting by a

molecular clock of mitochondrial DNA. J. Mol. Evol. 22: 160-174 (1985).

11) International Human Genome Sequencing Consortium. Initial sequencing and

analysis of the human genome. Nature 409: 860-921 (2001).

12) Kingman, J. F. C. Origins of the coalescent. Genetics 156: 1461-1463 (2000).

13) Kruglyak, L. et al. Parametric and nonparametric linkage analysis: A unified

approach. American Journal of Human Genetics 58: 1347-1363 (1996).

14) Kruglyak, L. Prospects for whole-genome linkage disequilibrium mapping of

common disease genes. Nature Genetics 22: 139-144 (1999).

15) Krushkal, J., et al. Genome-wide linkage analysis of systolic blood pressure

using highly discordant siblings. Circulation 11: 1407-1410 (1999).

16) Lai, E. et al. A 4-Mb high-density single nucleotide polymorphism-based map

around human APOE. Genomics 54: 31-38 (1998).

17) Lander, E.S. et al. Large-scale identification, mapping, and genotyping of

single nucleotide polymorphisms in the human genome. Science 280: 1077-

1082 (1998).

18) Li, W. Evolutionary change in nucleotide sequences. In: Li, W. editor.

Molecular Evolution. Sunderland: Sinauer Associates, Inc. p. 53 (1997).

19) McCarthy, J. J., and Hilfiker, R. The use of single-nucleotide polymorphism

maps in Pharmacogenomics. Nat. Biotech. 18: 505-508 (2000).

62

20) Nobukuni, Y. et al. Maple syrup urine disease: complete defect of the E1-beta

subunit of the branched chain alpha-ketoacid dehydrogenase complex due to a

deletion of an 11-bp repeat sequence which encodes a mitochondrial targeting

leader peptide in a family with the disease. J. Clin. Invest. 87: 1862-1866

(1991).

21) Risch, N. and Merikangas, K. The future of genetic studies of complex human

diseases. Science 273: 1516-1517 (1996).

22) Rothberg, B. E. G., Ramesh, T. M. and Burgess, C. E. Integrating expression-

based drug response and SNP-based pharmacogenetic strategies into a single

comprehensive pharmacogenomics program. Drug Development Research 49:

54-64 (2000).

23) Schafer, A. J., and Hawkins, J. R. DNA variation and future of human

genetics. Nature Biotechnology. 16: 33-39 (1998).

24) Schneider, S. et al. Arlequin: a software for population genetics data analysis

user manual version 2.0. Genetics and Biometry Lab. Department of

Anthropology, University of Geneva. (http://anthro.unige.ch/arlequin/).

25) Spielman, R. S., McGinnis, R. E. and Ewen, W. J. Transmission test for

linkage disequilibrium: the insulin gene region and insulin-dependent diabetes

mellitus (IDDM). Am J of Hum Genet. 52(3): 506-516 (1993).

26) Terwilliger, J. D., and Weiss, K. M. Linkage disequilibrium mapping of

complex diseases: fantasy or reality? Current opinion in Biotechnology 9:

578-594 (1998).

63

27) The International SNP Map Working Group. A map of human genome

sequence variation containing 1.42 million single nucleotide polymorphisms.

Nature 409: 928-933 (2001).

28) Venter, C. et al. The sequence of the human genome. Science 291(5507):

1304-1351 (2001).

29) Weir, B. Disequilibrium. In: Weir, B. S. editor. Genetic data analysis II -

methods for discrete population genetic data. Sunderland, MA: Sinauer

Associates. Inc. p. 91-139 (1996).

30) Zar, J. H. Testing for goodness of fit. In: Zar, J. H. editor. Biostatistical

analysis. New Jersey: Prentice Hall. p. 461-485 (1999) (a).

Zar, J. H. The normal distribution. In: Zar, J. H. editor. Biostatistical analysis.

New Jersey: Prentice Hall. p. 65-90 (1999) (b).

31) Zollner, S. and Haeseler, A. A Coalescent approach to study linkage

disequilibrium between single-nucleotide polymorphisms. Am. J. Hum.

Genet. 66: 615-628 (2000).

64

APPENDICES Appendix 1: An example of the SNP data for human chromosome 6 from the Whitehead Institute of Genome Research. In all there were 146 SNPs. (cR corresponds to the radiation hybrid distance). VNTR (variable number of tandem repeats) are the top and bottom markers for that particular SNP position. SNP NAME TYPE GENETIC

DISTANCE (cM) TOP VNTR BOTTOM

VNTR

WIAF 1583 A/G 1.4 D6S344 D6S344 WIAF 857 C/T 1.37(cR) D6S344 D6S344 WIAF 1034 C/T 0.00(cR) D6S344 D6S344 WIAF 2096 G/A 6.4 D6S1617 D6S1617 WIAF 131 T/C 41.75(cR) D6S1617 D6S1617 WIAF 132 T/C 41.75(cR) D6S1617 D6S1617 WIAF 950 A/C 51.37(cR) D6S296 D6S470 WIAF 549 C/T 9.0 D6S296 D6S470 WIAF 890 G/C 59.80(cR) D6S1674 D6S1674 WIAF 1939 C/G 54.83(cR) D6S1674 D6S1674 WIAF 1020 A/G 17.7 D6S470 D6S470 WIAF 881 G/A 67.56(cR) D6S470 D6S1578 WIAF 1567 G/A 20.5 D6S470 D6S1578 WIAF 967 T/C 66.11(cR) D6S470 D6S1578 WIAF 1891 C/T 81.31(cR) D6S469 D6S288 WIAF 116 G/A 85.33(cR) D6S469 D6S288 WIAF 1541 C/G 34.2 D6S1688 D6S1688 WIAF 1966 G/A 111.66(cR) D6S1688 D6S422 WIAF 709 T/A 34.0 D6S1688 D6S422 WIAF 105 C/T 118.48(cR) D6S422 D6S1686 WIAF 106 C/A 118.48(cR) D6S422 D6S1686 WIAF 415 T/C 121.60(cR) D6S1686 D6S1686 WIAF 1896 A/T 123.20(cR) D6S1686 D6S1691 WIAF 613 G/T 40.0 D6S1686 D6S1691 WIAF 614 A/G 40.0 D6S1686 D6S1691 WIAF 1959 T/C 128.64(cR) D6S1691 D6S464 WIAF 1574 T/C 130.56(cR) D6S464 D6S464 WIAF 1460 A/G 46 D6S276 D6S439 WIAF 1461 T/A 46 D6S276 D6S439 WIAF 1462 G/T 46 D6S276 D6S439 WIAF 2020 C/G 161.68(cR) D6S276 D6S439 WIAF 2021 T/C 161.68(cR) D6S276 D6S439 WIAF 139 A/C 165.13(cR) D6S276 D6S439 WIAF 1551 T/C 46.6 D6S276 D6S439 WIAF 1552 G/A 46.6 D6S276 D6S439 WIAF 1553 G/A 46.6 D6S276 D6S439 WIAF 1554 T/C 46.6 D6S276 D6S439 WIAF 1722 A/T 46.7 D6S276 D6S439

65

WIAF 110 C/T 179.84(cR) D6S276 D6S439 WIAF 556 C/T 177.90(cR) D6S276 D6S439 WIAF 557 C/T 177.90(cR) D6S276 D6S439 WIAF 558 T/C 177.90(cR) D6S276 D6S439 WIAF 257 C/A 178.21(cR) D6S276 D6S439 WIAF 258 A/C 178.21(cR) D6S276 D6S439 WIAF 1185 C/G 46.9 D6S276 D6S439 WIAF 1453 A/G 46.9 D6S276 D6S439 WIAF 1935 C/T 47.0 D6S276 D6S439 WIAF 897 C/T 178.21(cR) D6S276 D6S439 WIAF 898 A/G 178.21(cR) D6S276 D6S439 WIAF 1696 C/G 47.0 D6S276 D6S439 WIAF 643 A/G Unassigned D6S276 D6S439 WIAF 1306 A/G 47.1 D6S276 D6S439 WIAF 1084 C/T 47.2 D6S276 D6S439 WIAF 1009 T/C 47.2 D6S276 D6S439 WIAF 1335 G/C 179.74(cR) D6S276 D6S439 WIAF 1336 A/G 179.74(cR) D6S276 D6S439 WIAF 1337 C/A 179.74(cR) D6S276 D6S439 WIAF 1338 A/G 179.74(cR) D6S276 D6S439 WIAF 1339 G/A 179.74(cR) D6S276 D6S439 WIAF 1340 T/C 179.74(cR) D6S276 D6S439 WIAF 1341 G/A 179.74(cR) D6S276 D6S439 WIAF 1342 T/C 179.74(cR) D6S276 D6S439 WIAF 2177 G/C 47.7 D6S276 D6S439 WIAF 2106 C/T 47.7 D6S276 D6S439 WIAF 2103 A/G 47.8 D6S276 D6S439 WIAF 1006 T/C 188.58(cR) D6S276 D6S439 WIAF 1746 A/G 194.89(cR) D6S276 D6S439 WIAF 916 T/C 193.76(cR) D6S276 D6S439 WIAF 1586 G/A 52.8 D6S291 D6S1610 WIAF 1587 T/C 52.8 D6S291 D6S1610 WIAF 2075 G/A 65.5 D6S426 D6271 WIAF 1780 A/C 252.98(cR) D6S426 D6271 WIAF 1608 A/G 246.37(cR) D6271 D6271 WIAF 1609 A/G 246.37(cR) D6271 D6271 WIAF 1060 G/A 258.00(cR) D6271 D6271 WIAF 98 T/C 251.95(cR) D6271 D6S459

WIAF 1359 T/C 69.0 D6271 D6S459 WIAF 1360 T/C 69.0 D6271 D6S459 WIAF 538 C/T 69.0 TO 72.0 D6S459 D6S438 WIAF 539 C/A 69.0 TO 72.0 D6S459 D6S438 WIAF 997 G/T 271.85(cR) D6S459 D6S438 WIAF 520 C/T 77.0 D6S436 D6S466 WIAF 805 C/T 307.81 D6S466 D6S466 WIAF 2212 T/C 79.9 D6S257 D6S257 WIAF 589 T/A 462.24(cR) D6S1589 D6S1589

66

WIAF 922 A/C 533.70(cR) D6S1589 D6S1589 WIAF 1486 A/G 557.69(cR) D6S1601 D6S1570 WIAF 1669 G/A 557.21(cR) D6S1601 D6S1570 WIAF 1670 G/A 557.21(cR) D6S1601 D6S1570 WIAF 220 G/T 539.89(cR) D6S1601 D6S1570 WIAF 1654 A/C 102.4 D6S417 D6S424 WIAF 78 C/T 626.11(cR) D6S417 D6S424

WIAF 1427 G/C 618.65(cR) D6S417 D6S424 WIAF 1476 T/C 629.40(cR) D6S1716 D6S468 WIAF 207 G/A 116.2 D6S278 D6S278 WIAF 844 A/C 682.24 D6S278 D6S278 WIAF 845 A/C 682.24 D6S278 D6S278 WIAF 1875 C/T 116.6 D6S278 D6S302 WIAF 1876 G/A 116.6 D6S278 D6S302 WIAF 1877 G/C 116.6 D6S278 D6S302 WIAF 1878 T/C 116.6 D6S278 D6S302 WIAF 1879 C/T 116.6 D6S278 D6S302 WIAF 1880 A/G 116.6 D6S278 D6S302 WIAF 1881 C/T 116.6 D6S278 D6S302 WIAF 2219 T/G 116.6 D6S278 D6S302 WIAF 2126 G/A 116.9 D6S278 D6S302 WIAF 584 T/C 707.96(cR) D6S423 D6S423 WIAF 726 C/T 121 D6S423 D6S423 WIAF 985 T/A 705.43(cR) D6S423 D6S1712 WIAF 1203 G/T 121.0 D6S423 D6S1712 WIAF 2171 G/A 124.2 D6S423 D6S1712 WIAF 224 C/T 122.0(cR) D6S423 D6S1712 WIAF 225 T/C 122.0(cR) D6S423 D6S1712 WIAF 736 C/G 731.63(cR) D6S423 D6S1712 WIAF 737 G/A 731.63(cR) D6S423 D6S1712 WIAF 492 A/T 733.67(cR) D6S407 D6S262 WIAF 1674 T/C 736.71(cR) D6S407 D6S262 WIAF 738 A/G 129.0 D6S407 D6S262 WIAF 1366 G/A 744.83(cR) D6S407 D6S262 WIAF 642 T/C 138.0 D6S292 D6S1699 WIAF 1604 G/C 144.5 D6S453 D6S453 WIAF 466 A/G 748.63(cR) D6S453 D6S308 WIAF 1873 G/A 788.32(cR) D6S453 D6S308

WIAF 3 T/G 790.34(cR) D6S453 D6S308 WIAF 2034 T/C 145.6 D6S308 D6S308 WIAF 1704 T/G 150.4 D6S311 D6S1687 WIAF 1623 A/G 150.4 D6S311 D6S1687 WIAF 592 A/G 811.66(cR) D6S311 D6S1687 WIAF 982 C/T 814.38(cR) D6S1687 D6S1687 WIAF 563 A/C 155.0 D6S1687 D6S448 WIAF 762 C/G 837.32(cR) D6S419 D6S1579 WIAF 1954 T/C 165.7 D6S419 D6S1579 WIAF 1955 G/A 165.7 D6S419 D6S1579 WIAF 1493 T/A 165.7 D6S419 D6S1579

67

WIAF 1494 A/G 165.7 D6S419 D6S1579 WIAF 888 T/C 165.7 D6S419 D6S1579 WIAF 915 G/A 175.9 D6S1579 D6S1719 WIAF 97 G/A 833.43(cR) D6S1579 D6S1719

WIAF 397 C/T 843.63 D6S1719 D6S1719 WIAF 1690 A/C 178.4 D6S1719 D6S264 WIAF 30 C/T 846.82(cR) D6S1719 D6S264 WIAF 12 T/C 167.0 D6S1719 D6S264 WIAF 13 A/G 167.0 D6S1719 D6S264

WIAF 1739 A/G 184.8 D6S1585 D6S446 WIAF 1740 A/G 184.8 D6S1585 D6S446 WIAF 2186 A/G 189.0 D6S446 D6S1693

68

Appendix 2: The SNP data of human chromosome 6 from the Genome Database (PV = point variation) GENE

NAME SNP TYPE/ LOCATION CHANGE IN AMINO

ACID SNP

ABCB2 (TAP1)

PV Codon 333 ILE. to VAL.(ATC to GTC)

A/G

PV Codon 370 ALA to VAL (GCT to GTT)

C/T

PV Codon 458 VAL to LEU (GTG to TTG)

G/T

PV Codon 637 ASP to GLY(GAC

to GGC)

A/G

PV

Codon 648 ARG to GLN (CGA to CAA)

G/A

ABCB3 (TAP2)

PV Codon 163 VAL to VAL (GTC to GTT)

C/T

PV Codon 379 ILE to VAL (ATA to GTA)

A/G

PV Codon 386 GLY to GLY (GGG to GGT)

G/T

PV Codon 387 VAL to VAL (GTG to GTA)

G/A

PV Codon 436 ASN to ASN (AAC to AAT)

C/T

PV Codon 565 ALA to THR (GCT to ACT)

G/A

PV Codon 651 ARG to CYS (CGT to TGT)

C/T

PV Codon 665 ALA to THR (GCA to ACA)

G/A

PV Codon 687 GLN to STOP (CAG to TAG)

C/T

PV Codon 697 VAL to VAL (GTT to GTG)

T/G

C4A 4B Codon 1101 C to T C/T

Codon 1186 G to C G/C RFLP Exon

40-> nt 5095

CTG to CTA (no aa change)

G/A

CDKN1A PV Codon31 SER to ARG (C to A) C/A

PV Codon 91 A to T A/T

COL10A1 PV GLY to ARG PV2 exon 2 position 608 GTG to GTC G/C

69

CYP21A1P/

A2 PV exon 7

codon 269 SER to THR (G to C) G/C

ESR1 PV End of

Intron 1 and before exon2 (nt 257)

ALA to VAL (C to T) C/T

PV Nt 261 G to C (Silent) G/C

GMPR PV Nt 766 from chain initiat. Codon

PHE to ILE (T to A) T/A

PV C to T (silent) C/T

HSPA1B RFLP Nt 1267 A To G A/G

LPA PV Codon 4168 THR to MET (C to T) C/T

LTA/ TNFB (ii)

RFLP Intron 1 position 26 ASN to THR (AAC to ACC)

A/C

MEP1A PV Codon 176 A To G A/G

D6S282 To D6S272 A to G A/G

MLN PV Nt115 of gene Position-11of signal

peptide

VAL to ALA (T to C)

T/C

PV Nt 184 of 1st Exon C To G C/G

RDS PV Position 558 of gene C To T (no aa change) C/T

TNF (TNFA)

PV Nt –308 of 5' of the gene

G To A G/A

PV Nt –238 G To A G/A

70

Appendix 3: An example of the SNP data of human chromosome 6 from fifth release of the SNP Consortium.

SNP dbSNP # Chrom Band GenBank Version

Links Type of Change

TSC0089213 119642 Chr6 6p35.2 AL080251.3 a/g

TSC0089214 119643 Chr6 6p35.2 AL080251.3 c/t

TSC0016149 74583 Chr6 6q22.22 AL135839.2 a/g

TSC0148520 110255 Chr6 6q22.22 AL135839.2 a/g

TSC0089004 119501 Chr6 6q22.22 AL135839.2 a/g

TSC0010080 54655 Chr6 6q22.22 AL135839.2 a/g

TSC0026679 80938 Chr6 6q22.22 AL135839.2 a/g

TSC0097341 125463 Chr6 6p35.2 AL080251.3 g/t

TSC0119139 93589 Chr6 6q22.22 AL135839.2 c/g

TSC0102942 129620 Chr6 AC004842.2 NT_002179 g/t

TSC0110225 133601 Chr6 6p25 AL021328.1 NT_000213 a/c

TSC0110226 133602 Chr6 6p25 AL021328.1 NT_000213 a/c

TSC0090977 120849 Chr6 6p25 AL035696.14 NT_002179 a/g

TSC0067822 Chr6 6p25 AL035696.14 NT_002179 c/t

TSC0110227 133603 Chr6 6p25 AL021328.1 NT_000213 a/g

TSC0043505 59435 Chr6 6p25 AL035696.14 NT_002179 a/g

TSC0001995 25023 Chr6 6p25 AL035696.14 NT_002179 g/t

TSC0067821 Chr6 6p25 AL035696.14 NT_002179 a/g

TSC0102721 129487 Chr6 AC004842.2 NT_002179 a/g

TSC0069979 Chr6 6p25 AL021328.1 NT_000213 c/t Note: This example represents 20 SNPs out of 5102 SNPs collected from the SNP Consortium database for this project.